CN101187999A - Prediction system and prediction method for building fire smoke flow feature - Google Patents

Abstract

The invention requests for a system and a method for predicting fire fume characteristics to protect buildings and relates to predicting of fire fume development and evolution laws in high-rise buildings and multi-layer and multi-room buildings. The invention constitutes a complete building fire fume characteristic predicting system and a fire fume characteristic predicting method. A building is transformed into a building ventilation tree composed of nodes and branches so as to construct a junction matrix and a closed-loop matrix for ventilation circuits of the building; a fume characteristic equation is composed and solved to plot the relationship of time and concentration changes of CO2 and CO, then the time to reach the dangerous concentration is predicted. The system can predict the structure changes of the building during the development of the fire so as to make out forced ventilation proposals and smoke control and exhaust proposals for each part in the building and realize performance-based design of smoke prevention and exhaust systems and safety evacuation of the building with the maintenance of the fire safety.

Description

Building fire smoke flow feature prognoses system and Forecasting Methodology

Technical field

The invention belongs to fire safety evaluating and fire engineering field, be specifically related to the prediction of development of a kind of skyscraper and multilayer multicell building fire flue gas and development law.

Background technology

From the seventies in 20th century, some developed countries just begin the Building Fire Protection method for designing that turns to the basis with performance has been carried out systematic study in the world, occurred describing the various mathematics physics model of plume motion feature thereupon, mainly contained field model, regional model, network model and above-mentioned three kinds of models in conjunction with the mixture model of using.For skyscraper and multilayer multicell building thereof, because restricted clearance is a lot, boundary condition is very complicated, and the model of studying such building fire plume motion at present in the world adopts network model more.Mainly contain the BRI model of Japanese architecture research institute exploitation, the IRC model of Canadian Building Research Establishment exploitation, the BRE model of English architecture research institute exploitation, the NIST model of national institute of standards and technology's exploitation, the TNO model of Dutch applied physics research institute exploitation.From available data, all there is following point in above-mentioned model:

Promptly only with the function of flue gas concentration as the time, the physical parameter of the chamber that perhaps only will catch fire is as the function of time, and is considered as the temperature at other position in the buildings in the fire constant; The Calculation of Heat Transfer aspect has only been calculated the heat transfer of high-temperature flue gas to surface of wall, and has been ignored by the heat transfer of wall body to the adjacent room, therefore must cause the room temperature of adjacent room to calculate the generation deviation; With the thermal physical property parameter of flue gas, be reduced to constant as coefficient of heat conductivity, specific heat at constant pressure, density etc., this and actual conditions have very big difference, do not meet the rule of fire development and differentiation; Burning things which may cause a fire disaster heat release rate adopts the stable state burning things which may cause a fire disaster more in the fire, the stable state burning things which may cause a fire disaster can make and calculate simply, but its shortcoming is conspicuous, the setting power of stable state burning things which may cause a fire disaster often rule of thumb data determine, character is indeterminate, can not truly reflect the output situation of burning things which may cause a fire disaster heat release rate.And in common building fire, solid combustible accounts for the overwhelming majority, fire load density is very big, the output of its burning things which may cause a fire disaster heat release rate is that the evolution with fire changes, it is the function of time, adopt the stable state burning things which may cause a fire disaster as burning things which may cause a fire disaster heat release rate, very likely underestimate or over-evaluate fire risk, reduce the accuracy of result of calculation.

Summary of the invention

The present invention is directed to the above-mentioned problems in the prior art, set up complete building fire smoke flow feature model, propose, improve and finished theoretical calculation method, designed complete building fire smoke flow feature prognoses system.This system comprises: data center, control center, acquisition module, algoritic module, analysis and processing module,

(1) data center pre-deposits the relevant geometric data of building construction, the physical parameter of buildings; Acquisition module is gathered buildings environment weather condition;

(2) control center selects HRR to increase form, sets up the fire scenario model, in the fire generating process according to the change model that makes up burning things which may cause a fire disaster type, burning time fabric structure;

(3) burning time is selected the wall heat transfer submodel according to buildings by control center;

(4) control center is converted into the buildings ventilation tree that node and branch constitute according to constructure inner structure with buildings, constructs buildings ventilation loop continue matrix and closed-loop matrix thus;

(5) control center's setting-up time step delta τ according to buildings ventilation loop continue matrix, closed-loop matrix and correlation model and parameter, makes up the flue gas characteristic equation;

(6) algoritic module is found the solution the flue gas characteristic equation, calculates each node temperature, flue gas concentration, CO2 concentration, CO concentration, wall surface temperature, by flue gas mass rate, the pressure reduction of opening, generates above-mentioned each parameter curve over time;

(7) analysis and processing module is analyzed the time that flue gas arrives unsafe condition according to result of calculation, and CO concentration criterion, CO2 concentration criterion, temperature criterion, flue gas concentration criterion and flue gas layer height criterion are provided.

The present invention also proposes a kind of building fire smoke flow feature Forecasting Methodology, may further comprise the steps:

(1) the system data center pre-deposits the relevant geometric data of buildings, the physical parameter that buildings is relevant; Acquisition module is gathered buildings environment weather condition;

(2) select HRR to increase form, set up the fire scenario model, the change model of buildings structure in the fire generating process;

(3) select the wall heat transfer submodel burning time according to buildings;

(4) constructure inner structure is converted into the buildings ventilation tree that node and branch constitute, constructs buildings ventilation loop continue matrix and closed-loop matrix thus;

(5) setting-up time step delta τ according to buildings ventilation loop continue matrix, closed-loop matrix and correlation model and parameter, makes up the flue gas characteristic equation;

(6) find the solution each node temperature of flue gas characteristic Equation for Calculating, flue gas concentration, CO2 concentration, CO concentration, wall surface temperature, by flue gas mass rate, the pressure reduction of opening, generate each parameter curve over time;

(7) analyze the time that flue gas arrives unsafe condition according to result of calculation, and CO concentration criterion, CO2 concentration criterion, temperature criterion, flue gas concentration criterion and flue gas layer height criterion are provided.

With each room of buildings and staircase as node, stairwell is divided into different nodes by each floor face, corridor type passage is according to dividing into node with the approximately equalised area in room, the opening of passage is as branch, buildings is converted into the buildings ventilation tree that node and branch constitute, constructs buildings ventilation loop continue matrix and closed-loop matrix thus.

Adopt the building fire smoke flow feature prognoses system of the present invention's design, consideration is by the heat transfer of wall body to the adjacent room, can calculate in the fire process temperature of the temperature in each room of any time (node) in the buildings, smoke density, CO2 (carbon dioxide) concentration, CO (carbon monoxide) concentration, atmospheric density, room amount of stored heat, ventilation thermal loss, body of wall caloric receptivity, radiation heat loss, opening flue gas mass rate, opening pressure reduction, each node body of wall any time.Native system can be according to the various situations of predicting burning things which may cause a fire disaster type, burning time that building construction changes in the fire process, so that determine the pressure air blowing scheme and the anti-smoke evacuation design proposal at interior each position of buildings in the fire process, and the proposition suggestion for revision, realize guaranteeing that the performance chemoprevention smoke evacuation system under the building fire safety case designs.

Description of drawings

Fig. 1 flue gas characteristic Equation for Calculating of the present invention process flow diagram

Fig. 2 building fire smoke flow feature prognoses system of the present invention structural drawing

Fig. 3 empirical model plane and measuring point arrangenent diagram

Fig. 4 50kg, nothing smoke evacuation, 2 room temperature data comparison diagrams

Fig. 5 10kg, smoke evacuation is arranged, 1 room temperature data comparison diagram

Fig. 6 subterranean commercial area architectural plan

Fig. 7 subterranean commercial area network ventilation tree

Fig. 8 subterranean commercial area the 2nd and the 9th node CO ₂Concentration is schemed over time

Fig. 9 hotel building object plane figure

Figure 10 hotel building thing network ventilation tree

The contrast of Figure 11 hotel building thing 1 node temperature curve

The contrast of Figure 12 hotel building thing 1 node flue gas concentration curve

The contrast of Figure 13 hotel building thing 1 node CO2 concentration curve

Embodiment

Building fire smoke flow feature prognoses system of the present invention, set up complete building fire smoke flow feature model, the development and change rule of plume, cigarette temperature, smoke density, CO2 concentration, CO concentration and wall temperature in the time of can the predict good breaking out of fire, system also considers to change along with evacuating personnel the switching situation of each, causes coefficient of flow, the wall area of opening, the change of radiation transmission coefficient thus; Change the coefficient of flow, wall area, radiation transmission coefficient of opening etc. along with the door and window scaling loss.Model all is divided into node to all restricted clearances of buildings, can calculate plume rate of propagation and pollution range thereof, for personnel's safe escape dynamic similation provides Useful Information.

Below at accompanying drawing and instantiation enforcement of the present invention is elaborated, be illustrated in figure 2 as building fire smoke flow feature prognoses system model synoptic diagram of the present invention.

(1) in the system data central memory, pre-deposits physical parameters such as the coefficient of heat conductivity of coefficient of wind pres, coefficient of flow, radiation transmitance, each material of body of wall of relevant geometric data, the opening of the plane structure chart of buildings and three-dimensional structure diagram, building construction and volumetric ratio thermal capacitance; Acquisition module is gathered buildings environment weather condition by sensor, detector etc., comprises correlation parameters such as outdoor wind speed, blast, temperature;

(2) according to the functions of use of each chamber of buildings, set up the fire scenario model, select HRR to increase form, set up the fire scenario model that comprises stable state burning things which may cause a fire disaster and unstable state burning things which may cause a fire disaster.The stable state burning things which may cause a fire disaster comprises segmentation stable state burning things which may cause a fire disaster and piecewise linearity burning things which may cause a fire disaster; The unstable state burning things which may cause a fire disaster comprises, sets up square t that be directly proportional of HRR with temperature ²The unstable state burning things which may cause a fire disaster that burning things which may cause a fire disaster that is adopted among burning things which may cause a fire disaster, complete progressions model burning things which may cause a fire disaster, the multicell regional simulation software MRFC (Multi-Room-Fire-Code) that spreads in buildings according to fire and flue gas and FFB (fire research institute of Karlsruhe, Germany university) propose.Can select different burning things which may cause a fire disaster models for use according to each indoor combustible situation of buildings, and thus according to the burning things which may cause a fire disaster type and be configured in the change model of fabric structure in the fire generating process burning time;

(3), select the wall heat transfer submodel according to buildings length burning time of catching fire.Shorter as burning time, can select the big heat transfer submodel of semiinfinite, longer as burning time, can select limited thick heat transfer submodel.Latter's accuracy in computation is higher, but speed of convergence is slower.

(4) according to constructure inner structure its inner space and opening are divided into node and branch, and are converted into buildings ventilation tree, matrix [I] and closed-loop matrix [L] continue in formation buildings ventilation loop.With each room of buildings and each staircase as node, stairwell is divided into different nodes by each floor face, corridor type passage is according to dividing into node with the approximately equalised principle of room area, the opening of each room, passage is as branch, according to the graph theory principle buildings is converted into the buildings ventilation tree that node and branch constitute, forms buildings take a breath loop continue matrix [I] and closed-loop matrix [L] by buildings ventilation tree.

(5) system also can determine anti-smoke evacuation scheme according to fume exhausting type in fire process.As be the anti-smoke evacuation of machinery, then need import air output and exhaust smoke level, determine air output and exhaust smoke level according to the critical requirement of building construction variation and safe escape in the fire process, to satisfy the requirement of indoor occupant safe escape.

(6) system judges automatically whether buildings takes place with a loud crash.System is according to producing Hong the needed HRR I of combustion _c, the architectural openings area A _wWith open height H _w, all the other the room internal surface area A except that the architectural openings area _s, set up buildings discriminant with a loud crash according to Tomas Formula:

$I_c={387A}_w\sqrt{H_w}+{7.8A}_s\mathrm{kw}---\left(1\right)$

I _c≥425.52kw (2)

In the combustion process, can differentiate indoor Hong the combustion that taken place with formula (2) when satisfying formula (1) simultaneously.As Hong combustion does not take place, then calculate the indoor smoke flow feature that catches fire; As Hong combustion takes place, the chamber door and window that then catches fire crumbles, and flue gas flows into corridor or outdoor, calculates each node smoke flow feature.

(7) the computational accuracy setting-up time step delta τ as requested of system gets Δ τ=5m/s～10m/s usually, and Δ τ value I improves computational accuracy, but computing time is longer.According to the buildings ventilation loop matrix [I] that continues, closed-loop matrix [L], the change model of fabric structure, the wall heat transfer submodel, correlation parameter etc., utilize linearization and differencing to make up the flue gas characteristic equation, find the solution the flue gas characteristic equation according to the flue gas characteristic of current time, calculate next time step of each node temperature constantly, flue gas concentration, CO2 concentration, CO concentration, each room wall surface temperature of buildings etc. are according to aforementioned calculation value development and change rule in time, and the flue gas mass rate of opening, result of calculations such as pressure reduction are drawn and are generated each parameter curve over time.

(8) analysis module is analyzed the time that flue gas arrives unsafe condition according to change curve, and differentiates CO concentration, CO2 concentration, temperature, flue gas concentration and the flue gas layer height of unsafe condition automatically according to setting in advance threshold value in system database.Threshold value is to come out by a large amount of fire test data preparations, and deposits in the system database.

The concrete steps that make up described flue gas characteristic equation are as described below:

(1) according to buildings ventilation loop continue matrix and constant matrices, makes up node (room) mass balance equation formula.

If m node arranged in the buildings, n branch (opening), set up buildings ventilation tree, determine the buildings ventilation loop matrix [I] that continues according to ventilation tree, make up node and count the capable quality generation rate column matrix [M] of m that this matrix is a constant matrices, matrix continues in the ventilation loop that [I] is listed as for the capable n of m, be the matrix of coefficients of [Δ m], [Δ m] is the clean mass rate column matrix of the capable opening of n, is matrix to be asked.

Then node mass balance equation formula can be expressed as with matrix form:

[I][Δm]＝[M] (3)

(2) make up the loop pressure equilibrium equation according to buildings closed-loop matrix [L].

By the graph theory principle, can form n-m closed circuit by m node, the n buildings that branch constitutes a ventilation tree, can set up the algebraic equation of n-m linear independence, its matrix form is:

[L][P+P _s]＝[0] (4)

In the formula, [L] is that n-m is capable, n row closed-loop matrix, P _sBe pressure source; [P+Ps] is the capable internodal pressure reduction column matrix of n; [0] is 0 capable column matrix of n-m.

(3) according to architectural openings plane normal direction in the fire process and room, opening two ends smoke density, set up opening quality and opening pressure reduction relational expression.

Opening quality and opening pressure reduction relational expression are derived by fluid mechanics and are drawn, according to the plane of the opening normal direction, and room, opening two ends smoke density sets up different relational expressions, and the present invention adopts Taylor series expansion, ignore after the above infinite event of second order, make its linearity turn to following calculating formula:

(3.1) the plane of the opening normal vertically or room, opening two ends flue gas (air) density identical, be uniflux, opening quality of being set up and opening pressure reduction relational expression are:

$p_{n+1}=p_n-\frac{{\mathrm{\Δm}}_n^2}{\mathrm{\ρ}{\left(\mathrm{\μA}\right)}^2}+\frac{{\mathrm{\Δm}}_n}{\mathrm{\ρ}{\left(\mathrm{\μA}\right)}^2}{\mathrm{\Δm}}_{n+1}---\left(5\right)$

In the formula, p _N+1, p _nBe respectively (n+1) Δ τ moment and n Δ τ two internodal pressure reduction constantly; Δ m _N+1, Δ m _nBe respectively the clean mass rate of passing through the flue gas of opening between (n+1) Δ τ moment and the n Δ τ moment two nodes.μ is the opening flow coefficient, and A is an aperture area.

(3.2) plane of the opening normal level and room, opening two ends flue gas (air) density difference, the opening direction of establishing i chamber and j chamber points to the j chamber for just by the i chamber, and i chamber and j chamber have even temperature and density separately, and i chamber and j chamber density satisfy and concerns: ρ _i＜ρ _jAt this moment, neutral level should adopt different calculating formulas with respect to the diverse location of opening.

A, neutral level are in opening, and flue gas is two-way mobile mutually, and opening quality and opening pressure reduction relational expression are:

${\mathrm{\Δm}}_{n+1}={\mathrm{\Δm}}_n+\frac{{3\mathrm{Cp}}_n}{2D}(\sqrt{h_2-\frac{p_n}{D}}+R\sqrt{\frac{p_n}{D}-h_1})-\frac{3C}{2D}(\sqrt{h_2-\frac{p_n}{D}}+R\sqrt{\frac{p_n}{D}-h_1})p_{n+1}---\left(6\right)$

In the formula:

$C=\frac{2}{3}\mathrm{\&mu;<figure-callout\; id="2"\; label="B"\; filenames="S2007100929840E00041.png,S2007100929840E00051.png"\; state="\{\{state\}\}">B</figure-callout>}\sqrt{{2\mathrm{g\&rho;}}_i({\mathrm{\&rho;}}_j-{\mathrm{\&rho;}}_i)}---\left(7\right)$

$R=\sqrt{{\mathrm{\&rho;}}_j/{\mathrm{\&rho;}}_i}---\left(8\right)$

D＝(ρ _j-ρ _i)g (9)

$m_{i,j}=C{(h_2-\frac{p_n}{D})}^{3/2}---\left(10\right)$

m _j，i＝Δm _n-m _i，j (11)

h ₂, h ₁Be respectively the height of opening upper and lower side apart from ground, room; B is an opening degree; ρ _i, ρ _jBe respectively the indoor smoke density of i chamber and j; G is an acceleration of gravity; m _{I, j}, m _{J, i}Be respectively the i chamber and flow to the flue gas mass rate of j chamber and the mass rate that the j chamber flows to the i chamber.

B, opening are on neutral level, and flue gas is a uniflux, and opening quality and opening pressure reduction relational expression are:

${\mathrm{\&Delta;m}}_{n+1}={\mathrm{\&Delta;m}}_n-\frac{{3\mathrm{Cp}}_n}{2D}(\sqrt{h_1-\frac{p_n}{D}}-\sqrt{h_2-\frac{p_n}{D}})+\frac{3C}{2D}(\sqrt{h_1-\frac{p_n}{D}}-\sqrt{h_2-\frac{p_n}{D}})p_{n+1}---\left(12\right)$

C, opening are under neutral level, and flue gas is a uniflux, and opening quality and opening pressure reduction relational expression are:

${\mathrm{\&Delta;m}}_{n+1}={\mathrm{\&Delta;m}}_n+\frac{{3\mathrm{CRp}}_n}{2D}(\sqrt{\frac{p_n}{D}-h_1}-\sqrt{\frac{p_n}{D}-h_2})-\frac{3\mathrm{CR}}{2D}(\sqrt{\frac{p_n}{D}-h_1}-\sqrt{\frac{p_n}{D}-h_2})p_{n+1}---\left(13\right)$

If the branch that buildings is divided adds up to n, node adds up to m (not comprising the outdoor air node), and then the mass balance equation formula is m, and the pressure equilibrium equation is n-m, the linear equation of flow and pressure reduction is n, can form the Algebraic Equation set of 2n linear independence.The equation sum equals 2n unknown number number of opening two ends pressure reduction and opening flow composition, and the simultaneous solution system of equations can be obtained the pressure reduction and the opening mass rate at all opening two ends.

(4) set up the room heat balance equation according to real temperature, the following temperature T of solution node ^N+1

For arbitrary node i, adopt implicit difference scheme, the room heat balance equation can be rewritten as following finite difference equation:

$I_{\mathrm{ci}}=\underset{j}{\mathrm{\&Sigma;}}(C_{\mathrm{Pi}}^n{m}_{i,j}{T}_i^{n+1}-C_{\mathrm{Pj}}^n{m}_{j,i}{T}_j^{n+1})+\underset{l}{\mathrm{\&Sigma;}}{A}_{\mathrm{il}}{\mathrm{\&tau;}}_{\mathrm{il}}^n{\mathrm{\&sigma;}}_b(T_i^{n^3}{T}_i^{N+1}-T_l^{n^3}{T}_l^{n+1})---\left(14\right)$

$+\underset{k}{\mathrm{\&Sigma;}}{A}_{\mathrm{wi},k}{h}_{i,k}^n(T_i^{n+1}-T_{\mathrm{wi},k}^n)+{\mathrm{\&rho;}}_i^n{C}_{\mathrm{pi}}^n{V}_i\frac{T_i^{n+1}-T_i^n}{\mathrm{\&Delta;\&tau;}}$

In the formula: I _CiBe the chamber that catches fire (i chamber) combustible combustion heating speed, after the fire scenario of fire smoke flow feature model is set, I _CiCan calculate according to the burning things which may cause a fire disaster form of experimental data or default; T ⁿExpression τ=n Δ τ temperature constantly claims real temperature, is known quantity; T ^N+1Expression τ=(n+1) Δ τ temperature constantly claims following temperature, is amount to be asked; A _Il, τ _IlBe aperture area and the radiation transmitance between i, l two Room; σ _bBe Shi Difen-Boltzmann constant; A _{Wi, k}Be k the wall area in i chamber; h _{I, k}Be the total coefficient of heat transfer between k wall in i chamber and plume; T _{Wi, k}Be the surface temperature of k the wall in i chamber, try to achieve by the heat transfer equation formula; Smoke density ρ _i ⁿ, specific heat at constant pressure C _Pi ⁿDetermine Deng by current temperature; V _iBe i chamber volume.

(5) set up limited thick wall body heat transfer equation formula

Limited thick wall body heat transfer equation formula of the present invention is quoted the Yan Zhijun professor calculating formula in " Chongqing Univ. of Architecture's journal " 1997 the 5th interim paper of delivering " heat conduction of fire building is resolved ", but difference scheme is revised as implicit difference scheme by the demonstration difference scheme of original text.

(6) set up the big object heat transfer equation of semiinfinite formula according to total coefficient of heat transfer of plume and wall, the coefficient of heat conductivity and the thermal diffusivity of wall body material.

$T_{\mathrm{wi}\left(k\right)}^{n+1}=T_{\mathrm{wi}\left(k\right)}^n+{\mathrm{\&alpha;}}_{i\left(k\right)}^n(T_i^{n+1}-T_i^n)---\left(15\right)$

In the formula:

${\mathrm{\&alpha;}}_{i\left(k\right)}^n={\&Integral;}_0^{\mathrm{n\&Delta;\&tau;}}[\frac{h_{i\left(k\right)}}{{\mathrm{\&lambda;}}_{i\left(k\right)}}\sqrt{\frac{a_{i\left(k\right)}}{\mathrm{\&pi;\&theta;}}}-\frac{a_{i\left(k\right)}{h}_{i\left(k\right)}^2}{{\mathrm{\&lambda;}}_{i\left(k\right)}^2}\exp \left(\frac{a_{i\left(k\right)}{h}_{i\left(k\right)}^2}{{\mathrm{\&lambda;}}_{i\left(k\right)}^2}\mathrm{\&theta;}\right)\mathrm{erfc}\left(\frac{h_{i\left(k\right)}}{{\mathrm{\&lambda;}}_{i\left(k\right)}}\sqrt{a_{i\left(k\right)}\mathrm{\&theta;}}\right)]\mathrm{d\&theta;}---\left(16\right)$

In the formula, g is that integration intermediate variable h, λ, a are respectively total coefficient of heat transfer of plume and wall, the coefficient of heat conductivity and the thermal diffusivity of wall body material.In Matlab, α _{I (k)} ⁿCan come out with numerical calculations quickly and easily.When calculating the n+1 moment in step i chamber room temperature T _i ^N+1And k wall surface temperature T _{Wi (k)} ^N+1The time, n step T constantly _i ⁿ, T _{Wi (k)} ⁿ, α _{I (k)} ⁿBe given data, so T _{Wi (k)} ^N+1With T _i ^N+1Linear.

(7) set up the flue gas concentration equation, determine the flue gas mass concentration according to the cigarette generation rate that burning causes.

$(\frac{V_i{\mathrm{\&rho;}}_i^n}{\mathrm{\&Delta;\&tau;}}+\underset{j}{\mathrm{\&Sigma;}}{m}_{\mathrm{ij}})S_i^{n+1}-\underset{j}{\mathrm{\&Sigma;}}{m}_{\mathrm{ji}}{S}_j^{n+1}=M_{\mathrm{si}}+\frac{V_i{\mathrm{\&rho;}}_i^n}{\mathrm{\&Delta;\&tau;}}{S}_i^n---\left(17\right)$

In the formula: M _Si=c ₁I _Ci(18)

M _SiThe cigarette generation rate that causes for burning; c ₁The constant that the experiment of serving as reasons is determined; S _i, S _jBe respectively the flue gas mass concentration of i chamber and j chamber.

(8) set up CO ₂The concentration equation formula is determined the carbon dioxide mass concentration by the carbon dioxide generation rate that burning causes.

$(\frac{V_i{\mathrm{\&rho;}}_i^n}{\mathrm{\&Delta;\&tau;}}+\underset{j}{\mathrm{\&Sigma;}}{m}_{\mathrm{ij}})C_i^{n+1}-\underset{j}{\mathrm{\&Sigma;}}{m}_{\mathrm{ji}}{C}_j^{n+1}=M_{\mathrm{ci}}+\frac{V_i{\mathrm{\&rho;}}_i^n}{\mathrm{\&Delta;\&tau;}}{C}_i^n---\left(19\right)$

In the formula: M _Ci=c ₂I _Ci(20)

M _CiThe carbon dioxide generation rate that causes for burning; c ₂The constant that the experiment of serving as reasons is determined; C _i, C _jBe respectively the carbon dioxide mass concentration of i chamber and j chamber.

(9) CO concentration equation formula is determined the carbon monoxide mass concentration by the carbon monoxide generation rate that burning causes

$(\frac{V_i{\mathrm{\&rho;}}_i^n}{\mathrm{\&Delta;\&tau;}}+\underset{j}{\mathrm{\&Sigma;}}{m}_{\mathrm{ij}}){\mathrm{CO}}_i^{n+1}-\underset{j}{\mathrm{\&Sigma;}}{m}_{\mathrm{ji}}{\mathrm{CO}}_j^{n+1}=M_{\mathrm{coi}}+\frac{V_i{\mathrm{\&rho;}}_i^n}{\mathrm{\&Delta;\&tau;}}{\mathrm{CO}}_i^n---\left(21\right)$

In the formula: M _Coi=c ₃I _Ci(22)

M _CoiThe carbon monoxide generation rate that causes for burning; c ₃The constant that the experiment of serving as reasons is determined; CO _i, CO _jBe respectively the carbon monoxide mass concentration of i chamber and j chamber.

After setting up above flue gas characteristic equation, find the solution above-mentioned equation and determine correlation parameter, Figure 1 shows that concrete solution procedure.

(1) initial time is provided with n=0, supposes m between each node _{I, j}=0, pressure reduction p=0 obtains each node reality temperature T by formula (14) _i, obtain each wall surface temperature T by limited thick wall body heat transfer equation formula or the big object heat transfer equation of semiinfinite formula (16) _{W, i}

(2) put n=1, if n＜time step is carried out following steps;

(3) by the ρ of match _i=f (T _i), C _Pi= (T _i), with the n-1 moment each node smoke density ρ of temperature computation _i, specific heat at constant pressure c _PiObtain pressure reduction p between each node by perfect gas equation _i

(4) calculate n step branch flow, pressure reduction,, obtain the p of next time step by formula (3)-Shi (13) simultaneous solution _i ^N+1, Δ m _N+1, m _Ij ^N+1, m _Ji ^N+1

(5) step (4) is obtained next m _Ij ^N+1, m _Ji ^N+1Flue gas concentration, the CO of next time step are obtained in substitution formula (17), (19), (21) ₂Concentration and CO concentration;

(6) obtain the following temperature T of each node of next time step by formula (14) _i, and each wall surface temperature Twi;

(7) put n=n+1, repeating step (3) satisfies till the time that sets up to result of calculation.

Calculate flue gas concentration, the CO of each step-length ₂After the parameters such as concentration and CO concentration, draw the curve of above-mentioned parameter and time relationship, according to flue gas concentration, the CO that store in the database to human body generation unsafe condition ₂Concentration and CO concentration threshold compare, and determine to arrive the time of unsafe condition, and the flue gas concentration of current time, CO ₂Concentration and CO concentration.

With concrete buildings example enforcement of the present invention is described below.

Experimental study example 1 empirical model is an entity building, and high 3 meters, the layout of its planar dimension and measuring point as shown in Figure 3.Among Fig. 3, zero expression thermocouple assay post, △ are represented flue gas flow rate and smoke density, CO ₂Concentration collecting test post.In system database, import each opening physical dimension and fabric structure, material in advance, the coefficient of heat conductivity of each material of body of wall and volumetric ratio thermal capacitance, detector is gathered the environment weather condition.Burning things which may cause a fire disaster employing water percentage is 8%～10% long strip type pine, and its quality is respectively 10kg, 50kg, 100kg, has constituted three kinds of different fiery lotuses, and its combustion heating amount adopts weightless measurement method to measure.

Biao Zhu wall is not brickwork construction, plasters in the two sides.Wall 3 is a glass wall, and Room 3 are sight chamber, and 3 indoor searchlight and the camera systems of being provided with can be observed the plume motion conditions in the experiment.There is an exhaust opening at center, Room 2, can flue gas be discharged outdoor by this exhaust opening when blower fan is opened.Temp measuring system is that vertical 7 test poles arranging are formed, and respectively is provided with 9 temperature thermocouples on the test pole of Room 1 and Room 2, and the highest measuring point in position is apart from ceiling 0.1m, and down per two measuring point spacings are 0.2m.Respectively be provided with 4 temperature thermocouples on two test poles of Room 3, its height is respectively 0.1m, 0.3m, 1.3m, 2.55m apart from ceiling.The layout of test macro fully takes into account symmetry and representativeness.Measurement data is patrolled and examined collection automatically by data acquisition system (DAS), and polling period is 5 seconds, behind the data input computing machine of collection, by the printer real time print.Entire measuring device is advanced fire test isolated plant, and all thermopairs all pass through strict verification.

With combustion heating amount, exhaust smoke level and other primary data input system control center computing machines of actual measurement, getting time step is 5 seconds, adopts smoke flow feature prognoses system of the present invention to calculate.Result of calculation and experimental result are depicted on the same coordinate diagram, only provide the temperature data comparison diagram under representational two different schemes as space is limited, in this example.

From Fig. 4 and Fig. 5 as can be seen: under the less situation of fiery lotus (burning quality is respectively 50kg and 10kg), at the fire initial stage, the result of calculation of native system is a little more than measured result.Along with the carrying out of fire process, both errors reduce.When fiery lotus increased to 100kg, result of calculation and experimental data were coincide fairly goodly, and all comparative results have all fully confirmed this point.From network model each restricted clearance as a node, suppose that this node of arbitrary moment has the feature of physical parameters such as uniform temperature, density, when indoor combustible quantity was big, the forecasting software that adopts network model to develop had enough accuracys.Even fiery lotus only is 10kg, result of calculation also can be described the development and the variation tendency of building fire exactly.

Experimental study example 2 simulated object are that Sichuan fire-fighting scientific research is done and carried out the entity building that " subterranean commercial area Fire Smoke Flow experimental study " build.This building is an entity superstructure of building in 1: 1 ratio, and floor area of building is 640m ², head room is 3m, the long 67m of passage, wide 4m.A gateway is respectively established at two ends, and the passage both sides are the simulation paving.Its planimetric map as shown in Figure 6.

Following hypothesis has been done in this analog computation: (1) catch fire the chamber 1. the door standard-sized sheet, all paving complete shut-downs; (2) outlet 2 is closed, outlet 1,3 standard-sized sheets.Make up the ventilation tree of this buildings as shown in Figure 7.Buildings ventilation tree is a basic point 0 with outdoor, totally 10 node m=10, and 12 branch n=12, according to the graph theory principle, the ventilation loop that can set up m * n=10 * 12 matrix that continues, (n-m) * closed-loop matrix of n=2 * 12, totally 12 * 12 linear independence system of equations.

Experiment burning things which may cause a fire disaster condition: the chamber that catches fire is positioned at 1, and burning things which may cause a fire disaster is a long strip type timber, water percentage 8%～10%, cross section 5 * 5cm, length 25cm, quality 10kg.Determine the HRR of burning things which may cause a fire disaster during burning with weight-loss method.The opening condition: the paving complete shut-down, outlet 2 is closed, and outlet 1,3 is opened.Anti-fume exhausting type: mend wind, natural draught system naturally.Outdoor conditions: calm, 20 ℃ of temperature.

Test macro: " ● " locates to arrange the combined test post in six cross sections of P1～P6 in Fig. 6 corridor, each test pole upper edge short transverse is furnished with nine temperature points, and its height is respectively 0.1m, 0.3m, 0.5m, 0.7m, 0.9m, 1.1m, 1.3m, 1.5m, 2.55m apart from the furred ceiling distance; Three flow velocity measuring points, its height is respectively 0.5m, 1.0m, 1.5m apart from the furred ceiling distance; Six smoke density measuring points, its height is respectively 0.1m, 0.3m, 0.6m, 0.9m, 1.2m, 1.5m apart from the furred ceiling distance; A chemical analysis gas production sampling point, its height is 1.5m apart from the furred ceiling distance.

Carry out in the process of analog computation at utilization native system forecast model, getting time step is 5 seconds, and simulated time is 1800 seconds totally 360 steps.

To comprising outdoor node (1. ... 10.) about 10 minutes consuming time of the simulation of totally 11 nodes.Adopt the volume weighting method of average to ask for node temperature mean value, algoritic module calls constructed flue gas characteristic equation and carries out correlation computations.Result of calculation is consistent with the experimental result variation tendency, P1 test pole experimental data and 2 node computational datas are listed in the table 1, contrast as seen from table 1: temperature maximal value relative error is no more than 14%, temperature-averaging value relative error is no more than 3% in full-time the scope, simulation is successful to the proof model to the corridor flue-gas temperature, and Simulation result is reliable.Other node places also can draw identical conclusion.Because what adopt when calculating HRR is weight-loss method, all weight that lose are all considered to be converted into heat, and timber contains 8%～10% moisture content.Before burning entered the stabilization sub stage, water evaporates not only can not be generated heat, and also will absorb the latent heat of vaporization, thus higher according to the HRR of Model Calculation, temperature maximal value time of occurrence inclined to one side morning.After fire entered the smooth combustion stage, two curves reached unanimity.

Table 1P1 temperature experiment data and computational data are relatively

The rate of propagation of flue gas in the corridor: flue gas rate of propagation and distance in the horizontal direction is the important evidence of smoke control system design and safe escape design, is the important indicator that can differentiation fire initial stage personnel safe escape.Flue gas concentration, CO ₂Concentration and temperature variation all can reflect the situation that spreads of flue gas, in this analog computation, get CO ₂The situation of change of concentration is analyzed.By certain node CO ₂The time that concentration changes represents that flue gas invades the time of node, relatively each node flue gas concentration, invade morning and evening and the euclidean distance between node pair of time, can draw the level of flue gas in the corridor and spread average velocity, thereby provide foundation for safe escape design.

Fig. 8 is the 2nd and the 9th node CO ₂Concentration is schemed over time, and as can be seen from Figure: the time that the Model Calculation flue gas is invaded node 2 is 40 seconds, and the time of invading node 9 is 180 seconds, and the distance between node 2 and the node 9 is 45.8m.Therefore, the average rate of propagation of the level of flue gas in the corridor is: v _c=45.8/ (180-40)=0.33m/s.The P1 place is apart from furred ceiling first height H in the experiment ₁It is 75 seconds that the cigarette induction device of=0.1m is experienced the flue gas invasion time, and the P6 place is apart from furred ceiling first height H ₁It is 210 seconds that the cigarette induction device of=0.1m is experienced the flue gas invasion time, and the average rate of propagation of level is: v _e=45.8/ (210-75)=0.34m/s, the two relative error is 2.9%.It is very little that the level of flue gas in the corridor spreads the average velocity error.The flue gas intrusion time has determined the scope of a certain moment smoke pollution, and result of calculation wants Zao than the experimental result intrusion time, this be because: cigarette induction device itself is to the response existence hysteresis of flue gas.Flue gas arrives after the cigarette induction device point position, and the cigarette induction device just can be made reflection behind certain hour.For the safe escape design, the gained result is content to retain sovereignty over a part of the country complete like this.

This example is used building fire flow of flue gas characteristic prognoses system, parameters such as the flue-gas temperature of the corridor type passage under the little fiery lotus condition of 10kg timber, concentration are simulated, these parameter distributions situations have been drawn, and contrast with Sichuan Province's fire science research institute of Ministry of Public Security entity fire test data, comparing result shows: prognoses system is temperature, CO in the passage during to corridor type passage breaking out of fire ₂Predict the outcome and the experimental result of the average rate of propagation of level of CONCENTRATION DISTRIBUTION situation and flue gas are coincide very goodly, thereby prove that this software is reliable to flue gas predicting the outcome of flow characteristics in the type passage of corridor

Certain layer (as shown in Figure 9) of application example hotel building thing, totally 10 guest rooms, each room is connected with the corridor by singly opening wooden door, and the corridor does not have direct natural ventilation exterior window, and personnel enter the staircase cup by the fire-proof door of corridor right-hand member during fire.Be ventilation tree, the ventilation tree that Figure 10 obtains for the simulant building thing is abstract with buildings is abstract.If No. 3 room of simulant building thing catches fire (because of No. 3 the most close outlet in room, it catches fire and is the least favorable condition), common finishing, burning things which may cause a fire disaster HRR are 125KW/m ², suppose that burning things which may cause a fire disaster is the unstable state burning things which may cause a fire disaster, HRR is with t ²The fire form increases, and propagation of flame is a middling speed fire development form, and the fire development coefficient value is 0.0117KJ/s ³, maximum heat rate of release value is 3100kw.Computing time, step-length was 5 seconds, and calculating step number was 360 steps.

Adopt the negative pressure mechanical smoke extraction.The smoke evacuation area calculates by the interior aisle smoke evacuation area of " high rule " regulation, i.e. the floor area 40.5m in aisle ²Area 24.75m with the room 3 of the fixed window that is communicated with the aisle ²Sum, the area of always discharging fume is 65.25m ², the mechanical smoke extraction mouth is located at node 2 places, middle part, corridor, " when bearing a smoke bay smoke evacuation, should be not less than 60m by every square metre of area with existing " high rule " ³/ h calculates " be reference point, exhaust smoke level is expanded to the left and right sides, with 6m ³/ h.m ²Be a step-length, calculate 54m respectively ³/ h.m ², 60m ³/ h.m ², 66m ³/ h.m ², 72m ³/ h.m ², 78m ³/ h.m ²Exhaust smoke level the time flow of flue gas characteristic during building fire.

Under the situation that least favorable situation when building fire is caught fire in No. 3 room of the most close outlet, each room personnel will realize safe escape, must guarantee that the required safe escape times of each room personnel (RSET) are less than its available safe escape time (ASET).After fire took place, the guest can only enter the escape of staircase cup by the right fire-proof door in the room.Installing smoke warning device in the hotel, back 60s takes place and reports to the police in fire; The action speed of personnel in the room is 1.0m/s, evacuation speed 0.5m/s in the corridor; Confirm reaction time 90s, calculate this and build the maximum required safe escape time (from No. 13 required safe escape in the room time in outlet room farthest) REST=262s.

The corridor is to the fire-proof door of staircase cup standard-sized sheet preceding 60 seconds the time during natural draught system, and the right fire-proof door is closed after 60 seconds; The window of chamber of catching fire was closed at preceding 300 seconds, burst apart in the time of 300 seconds; Set No. 10, No. 13, No. 8, No. 6 windows at random and open, total open area is 10.8m ², all the other windows are closed.Utilization plume prognoses system model is simulated with the flow of flue gas characteristic under the different mechanical smoke extraction amount conditions natural draught system respectively, in order to save space, only draws the flue gas characteristic comparison diagram of No. 1 node in this example.

Contrast as can be seen from Figure 11-13: during natural draught system, each node temperature, flue gas concentration all are the trend of continuous rising; When mechanical smoke extraction was set, each node flue gas concentration was ascendant trend comparatively fast earlier, and ascendant trend is slowed down after the arrival certain hour, tends towards stability.The temperature of each node, smoke density all are lower than the situation with moment natural draught system, and exhaust smoke level is big more, and it is big more to descend.The setting that shows smoke evacuation system has delayed the rising of flue-gas temperature, concentration, and vital role has been played in escape to safe escape.

Relatively fire hazard state criterion (establish critical temperature and be 78 ℃, critical flue gas massfraction and be 0.053, critical CO ₂Massfraction is 0.046), correlation curve figure can see under the natural draught system situation, No. 1 node temperature value when 52 steps were 260s is 97.4 ℃, is much higher than the precarious position critical value, can not satisfy the safe escape requirement.In exhaust smoke level is 60m ³/ (h.m ²) time, temperature value was 76.4 ℃ when No. 1 node in corridor was 260s in 52 steps, temperature value was 78.3 ℃ when 53 steps were 265s, was lower than the temperature threshold value of this buildings under simulated conditions.This moment, flue gas concentration was 0.042kg/m ³, CO ₂Concentration is 0.023kg/m ³, all be lower than the critical value of fire hazard state.Be AEST＞REST, visible common finishing, burning things which may cause a fire disaster HRR are 125KW/m ²Condition under, the mechanical smoke extraction amount is pressed 60m ³/ (h.m ²) design is fully feasible.But total exhaust smoke level L=65.25 * 60=3915m ³/ h is less than the minimum exhaust smoke level L of " high rule " regulation _Min=7200m ³/ h.

Along with exhaust smoke level increases to 72m ³/ (h.m ²), smoke discharging effect is than 60m ³/ (h.m ²) be significantly improved, further increase exhaust smoke level to 78m ³/ (h.m ²), no matter be temperature, or smoke density or CO ₂Concentration is all with 72m ³/ (h.m ²) time very approaching, smoke discharging effect has not had obvious raising, its reason is that exhaust smoke level increases to a certain degree, indoor negative pressure increases, smoke exhaust efficiency descends, and causes smoke discharging effect not have obvious raising.Therefore it is unhelpful increasing exhaust smoke level greatly excessively.

For different fire location situations, because of the chamber 3 the most close outlets of catching fire, safe escape least favorable situation when being fire is the situation when so the chamber that do not need to consider again to catch fire is positioned at other positions.

Constant when other conditions, change burning things which may cause a fire disaster intensity, establishing above-mentioned buildings is fine fitment, HRR is 250kw/m ², promptly maximum heat rate of release value is 6200kw, when increasing 1 times than common finishing.With the exhaust smoke level is 72m3/ (h.m ²) calculate, can draw node temperature curve, node smoke density curve, node CO according to result of calculation ₂Concentration curve.Analyze according to curve, can determine the best smoke evacuation scheme of this buildings.

This example as protected object, has rationally proposed fire smoke precarious position criterion with the personnel in the fire building, determines to provide necessary condition for ASET.Putting before this, analyzing the mechanical smoke extraction amount, result of calculation shows: fit up for general when buildings, burning things which may cause a fire disaster intensity is (125KW/m not too greatly ²) time, 60m ³/ (h.m ²) the critical requirement of exhaust smoke level safe escape can satisfy fire the time; When buildings is fine fitment, the big (250KW/m of burning things which may cause a fire disaster intensity ²) time, 60m ³/ (h.m ²) exhaust smoke level can not satisfy safety evacuation requirement, 72m ³/ (h.m ²) exhaust smoke level can meet the demands, and it is not remarkable further to increase under the situation of exhaust smoke level actual effect.So 72m ³/ (h.m ²) exhaust smoke level be this moment the optimal mechanical exhaust smoke level.This result and Sichuan fire science research institute match by the conclusion that real experiment drew.The important conclusion of another that obtains according to the present invention is: total exhaust smoke level=unit area exhaust smoke level * smoke evacuation area, the minimum exhaust smoke level L that not stipulated by " high rule " _Min=7200m ³The restriction of/h.

The forecast model that employing the present invention sets up has important references value to the performance design of buildings smoke control system.

Claims (10)

1. building fire smoke flow feature prognoses system, this system comprises: data center, control center, acquisition module, algoritic module, analysis and processing module is characterized in that: (1) data center pre-deposits the relevant geometric data of building construction, the physical parameter of buildings; Acquisition module is gathered buildings environment weather condition;

(2) control center selects HRR to increase form, sets up the fire scenario model, according to the burning things which may cause a fire disaster type and set up the change model of fabric structure burning time;

(3) burning time is selected the wall heat transfer submodel according to buildings by control center;

(4) control center is converted into the buildings ventilation tree that node and branch constitute according to constructure inner structure with buildings, constructs buildings ventilation loop continue matrix and closed-loop matrix thus;

(5) control center's setting-up time step delta τ according to buildings ventilation loop continue matrix, closed-loop matrix and correlation model and parameter, makes up the flue gas characteristic equation;

(6) algoritic module is found the solution the flue gas characteristic equation, calculates each node temperature, flue gas concentration, CO2 concentration, CO concentration, wall surface temperature, by flue gas mass rate, the pressure reduction of opening, generates flue gas concentration, CO2 concentration, CO concentration curve over time;

(7) analysis and processing module is analyzed the time that flue gas arrives unsafe condition according to result of calculation, and CO concentration criterion, CO2 concentration criterion, flue gas concentration criterion and flue gas layer height criterion are provided.

2. building fire smoke flow feature prognoses system according to claim 1 is characterized in that, described buildings ventilation tree comprises the node that each room, each staircase, each the floor face by buildings constitutes, the branch that is made of access portal.

3. building fire smoke flow feature prognoses system according to claim 1 is characterized in that, described physical parameter comprises: the coefficient of heat conductivity and the specific heat capacity of the coefficient of wind pres of architectural openings, coefficient of flow, radiation transmitance, each material of body of wall.

4. building fire smoke flow feature prognoses system according to claim 1 is characterized in that, system also determines air output and exhaust smoke level according to the critical requirement of fume exhausting type, building construction variation, safe escape.

5. according to one of them described building fire smoke flow feature prognoses system of claim 1-4, it is characterized in that the fire scenario model of being set up comprises, stable state burning things which may cause a fire disaster and unstable state burning things which may cause a fire disaster model.

6. according to the described building fire smoke flow feature prognoses system of claim 1-4, it is characterized in that, the described flue gas characteristic equation that makes up comprises: according to buildings ventilation loop continue matrix and constant matrices, make up node mass balance equation formula, make up the loop pressure equilibrium equation according to the buildings closed-loop matrix; According to architectural openings plane normal and room, opening two ends smoke density, the buildings neutral level is set up opening quality and opening pressure reduction relational expression with respect to the diverse location of opening; Set up the room heat balance equation according to real temperature; Set up limited thick wall body heat transfer equation formula; Set up the big object heat transfer equation of semiinfinite formula according to total coefficient of heat transfer, the wall body material of plume and wall; Set up flue gas concentration equation, CO2 concentration equation formula, CO concentration equation formula.

7. a building fire smoke flow feature Forecasting Methodology is characterized in that, may further comprise the steps:

(1) the system data center pre-deposits the relevant geometric data of buildings, the physical parameter that buildings is relevant, and acquisition module is gathered buildings environment weather condition;

(2) select HRR to increase form, set up the fire scenario model, according to the burning things which may cause a fire disaster type and set up the change model of fabric structure burning time;

(3) select the wall heat transfer submodel burning time according to buildings;

(4) constructure inner structure is converted into the buildings ventilation tree that node and branch constitute, constructs buildings ventilation loop continue matrix and closed-loop matrix thus;

(5) setting-up time step delta τ according to buildings ventilation loop continue matrix, closed-loop matrix and correlation model and parameter, makes up the flue gas characteristic equation;

(6) find the solution each node temperature of flue gas characteristic Equation for Calculating, flue gas concentration, CO2 concentration, CO concentration, wall surface temperature, by flue gas mass rate, the pressure reduction of opening, generate each parameter curve over time;

(7) arrive the time of unsafe condition according to flue gas characteristic Equation for Calculating interpretation of result flue gas, and CO concentration criterion, CO2 concentration criterion, temperature criterion, flue gas concentration criterion and flue gas layer height criterion are provided.

8. building fire smoke flow feature Forecasting Methodology according to claim 7, it is characterized in that, with each room of buildings and staircase as node, stairwell is divided into different nodes by each floor face, corridor type passage is according to dividing into node with the approximately equalised principle of room area, the opening of passage is converted into buildings the buildings ventilation tree of node and branch formation as branch.

9. building fire smoke flow feature Forecasting Methodology according to claim 7 is characterized in that, the fire scenario model of being set up comprises, two kinds of fire scenario models of stable state burning things which may cause a fire disaster and unstable state burning things which may cause a fire disaster.

10. building fire smoke flow feature Forecasting Methodology according to claim 7, it is characterized in that, the step of described structure flue gas characteristic equation comprises: according to buildings ventilation loop continue matrix and constant matrices, make up node mass balance equation formula, make up the loop pressure equilibrium equation according to the buildings closed-loop matrix; According to architectural openings plane normal and room, opening two ends smoke density, set up opening quality and opening pressure reduction relational expression; Set up the room heat balance equation according to real temperature; Set up limited thick wall body heat transfer equation formula; Set up the big object heat transfer equation of semiinfinite formula according to total coefficient of heat transfer of plume and wall, the coefficient of heat conductivity and the thermal diffusivity of wall body material; Set up flue gas concentration equation, CO2 concentration equation formula, CO concentration equation formula.

Images